1. Field of the Art
The present invention relates to a YC separator for a video signal processing circuit, and more specifically to a separator for separating a Y (luminance) signal and a C (carrier chrominance) signal from composite video signals in video cassette recorders, for instance.
2. Prior Art
The same applicant has already proposed Video Signal Processing Circuit in Japanese Patent Application No. 62-140921, which comprises, as shown in FIG. 1, a first circuit (bandpass filter) 12 for separating a composite video signal to obtain a first chrominance signal a including part of luminance signal components; a second circuit 10 for generating a second chrominance signal c' including a chrominace separation error signal by removing the luminance signal components from the first chrominance signal a; and a third circuit 11 for generating a chrominance signal c including no chrominance separation error signal by selecting the second highest potential signal from among the first chrominance signal a, the second color signal c' and a reference potential.
In FIG. 2, the reference numeral 2 denotes a 1 H delay circuit; 23 denotes a .DELTA.t delay circuit; 17 and 22 denote an adder, respectively; and 8 denotes a subtractor. Further, the numerals 13, 15, 18 and 21 denote MAX circuits or higher potential detectors, respectively, each of which is formed as shown in FIG. 2(A) to output one signal with a higher potential from between two input signals. The numerals 14, 16, 19 and 20 denote MIN circuits or lower potential detectors, respectively, each of which is formed as shown in FIG. 2(B) to output one signal with a lower potential from between two input signals.
A truth table of a circuit shown in FIG. 1 is listed in Table 1 below:
TABLE 1 ______________________________________ a b c' c(C) Y ______________________________________ 0 0 0 0 0 1 0 1 1 0* 0 1 -1 0 0* 1 1 0 0 1** 1 -1 1 1 0 -1 1 -1 -1 0 ______________________________________
In Table 1 above, since C=1 or 0 at positions indicated by mark *, it is possible to improve the vertical resolution power required for small characters, for instance as compared with a prior-art comb filter where C=1/2, -1/2. This is because it is possible to eliminate half-tone or discoloration of the C signals generated at vertical transition portion of color bar signal or dot crawl generated at the Y signal due to the C signal.
In the prior-art circuit shown in FIG. 1, when magenta is displayed from the uppermost line to the (n-1)th line and green is displayed from the n-th line to the lowermost line on a picture as shown in FIG. 3(A), since C signals at two adjacent lines are opposite polarity to each other, for instance at the (n-2)th line and the (n-1)th line, a vertical correlation exists. However, since C signals are the same polarity, for instance of the (n-1)th line and the n-th line, no vertical correlation exists, as depicted in FIG. 3(B). In particular, at the border between the (n-1)th line magenta and the n-th line green (at which no vertical correlation exists), since a=1, b=1, C=0, and Y=1 as listed at position indicated by mark ** in Table 1, the C signal is apparently the same as the highband Y signal (a vertical correlation exists in Y signal).
In other words, all the C signal energy enters the Y signal side at the current n-th line as indicated by mark ** in Table 1. As a result, dot crawl occurs at n-th line as shown in FIG. 5, thus deteriorating the picture quality.
On the other hand, FIG. 6 shows another prior-art YC separator. In this circuit, a composite video signal (e.g. color bar signal) applied to a terminal 101 is passed through a bandpass filter 102 and a filter circuit 103 (described later in further detail) to generate a Cc (carrier chrominance) signal through a terminal 104. On the other hand, the composite video signal is supplied to an adder 107 via a .DELTA.t delay circuit 105 and a 1 H delay circuit 106, and then added to the Cc signal to generate a Yc (luminance) signal through a terminal 108. This YC separator utilizes a vertical correlation of video signals.
In the prior-art comb filter utilizing the vertical correlation, two-line vertical correlation is usually obtained on the basis of the present line information and the preceding (1 H past) line information. In this circuit shown in FIG. 6, however, the correlation is predicted by three lines of the present, 1 H past and 2 H past line information.
In the filter circuit 103, the above-mentioned three line information can be obtained by three signals A, B, and C, where signal A is an input signal (indicative of future), signal B is an output signal of a 1 H delay circuit 109 (indicative of present) and signal C is an output signal of another 1 H delay circuit 110 (indicative of past). In FIG. 6, the reference numerals 111, 112 and 113 denote higher potential detectors, respectively (referred to as MAX circuits), each of which outputs one higher potential signal of two input signals. The reference numerals 114, 115 and 116 denote lower potential detectors, respectively (referred to as MIN circuits) each of which outputs one lower potential signal of two input signals.
For example, when video signals of NTSC system are classified by selecting three lines when seen along the vertical direction in the picture, it is possible to obtain roughly three patterns as shown in FIGS. 7(A), (B) and (C). FIG. 7(A) shows a flat pattern; FIG. 7(B) shows a step pattern; and FIG. 7(C) shows a pulse pattern. Here, n denotes a current point on any given raster; (n-1) denotes a past point on the raster; and (n+1) denotes a future point on the raster. In the case of Cc signal, for instance, since the frequency of the subcarrier signal is fsc=(455/2)fH where fH denotes the horizontal scanning frequency, when the vertical correlation exists, it is possible to obtain a pulse pattern as shown in FIG. 7(C), where the lines change alternately. This pulse pattern can be obtained by only the three line information.
FIG. 8 is a block diagram for assistance in explaining the basic operation of the filter circuit 103. The MAX circuit 111 outputs one higher potential signal A or B of the two signals A and B; the MAX circuit 112 outputs one higher potential signal B or C of the two signals B and C; the MIN circuit 114 outputs a lower potential output X.sub.(+) of the two outputs of the two MAX circuits 111 and 112, which can be expressed as EQU X.sub.(+) =MIN (B, MAX (A, C)).
In the same way, the MIN circuit 115 outputs a lower potential signal A or B of the two signals A and B; the MIN circuit 116 outputs a lower potential signal B or C of the two signals B and C; the MAX circuit 113 outputs a higher potential output X.sub.(-) of the two outputs of the two MIN circuits 115 and 116, which can be expressed as EQU X.sub.(-) =MAX (B, MIN(A, C)).
These two signals X.sub.(+), X.sub.(-) are added by an adder 117. The level of the added signal is reduced to a half via a 1/2 circuit 118 and outputted as a chrominance signal Cc=1/2 X.sub.(+) +X.sub.(-)). Here, in this circuit, the present line signal level is inverted for calculation in order to obtain the Cc signal earlier.
In this case, the transfer function of the filter circuit 103 can be expressed as EQU C.sub.comb =1/2(B+MID(A,B,C))
where MID(A,B,C) is a function to output the second highest data of three input signals A, B, and C. Therefore, the circuit shown in FIG. 8 generates four Cc signals according to four patterns.
In the prior-art circuit shown in FIG. 6, when a picture of vertical stripes (in which black and white lines are alternately arranged such as multiburst) as shown in FIG. 9(A) is displayed, there exists a problem in that cross color (originally colorless position is colored) is produced at the upper and lower ends of a picture as shown in FIG. 9(B) and the Yc signal blur occurs. In more detail, when there exist vertical stripes below the position B as shown in FIG. 9(A), the three line information of the upper end is A=0, B=C=1 as shown in FIG. 10(A). Since B is inverted to form the Cc signal, the information changes to A=0, B=-1, C=1. Therefore, (-B+A)/2=1/2 is extracted from the transfer function=1/2 (B+MID(A, B, C)) as the Cc signal, so that there exists a problem (cross color) in that originally colorless positions are colored. On the other hand, there exists another problem in that blur occurs in the Yc signal because the amplitude thereof is reduced to half as Yc= 1-1/2=1/2.